We analysed the data from the control group in a typhoid vaccine trial in Karachi to assess the differences in individual-, household- and cluster-level characteristics for developing typhoid fever. The annual incidence of typhoid in children aged 2–16 years in the control arm of the vaccine trial was 151/100 000 population. After adjustment, the risk of typhoid was lower with increasing age [risk ratio (RR) 0·89, 95% confidence interval (CI) 0·83–0·95], was higher with an increase in population density (RR 1·13, 95% CI 1·05–1·21) and was lower in the households using a safe drinking-water source (RR 0·63, 95% CI 0·41–0·99). Typhoid fever affects younger children living in areas of high population density and lack of access to safe water in Pakistan. A combination of environmental and biological interventions is required to prevent the continued epidemiological and economic impact of typhoid fever in high-risk areas of Pakistan.